What time is it? Temporal Analysis of Novels

Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing, pages 9076–9086, November 16–20, 2020. c⃝2020 Association for Computational Linguistics 9076 What time is it? Temporal Analysis of Novels Allen Kim, Charuta Pethe, Steven Skiena Department of Computer Science, Stony Brook University, NY, USA {allekim,cpethe,skiena}@cs.stonybrook.edu Abstract Recognizing the flow of time in a story is a crucial aspect of understanding it. Prior work related to time has primarily focused on identi- fying temporal expressions or relative sequenc- ing of events, but here we propose computa- tionally annotating each line of a book with wall clock times, even in the absence of ex- plicit time-descriptive phrases. To do so, we construct a data set of hourly time phrases from 52,183 fictional books. We then con- struct a time-of-day classification model that achieves an average error of 2.27 hours. Fur- thermore, we show that by analyzing a book in whole using dynamic programming of break- points, we can roughly partition a book into segments that each correspond to a particu- lar time-of-day. This approach improves upon baselines by over two hour. Finally, we apply our model to a corpus of literature categorized by different periods in history, to show inter- esting trends of hourly activity throughout the past. Among several observations we find that the fraction of events taking place past 10 P.M jumps past 1880 - coincident with the advent of the electric light bulb and city lights. 1 Introduction The flow of time is an indispensable guide for our actions, and provides a framework in which to see a logical progression of events. Just as in real life, the clock provides the background against which literary works play out: when characters wake, eat, and act. In most works of fiction, the events of the story take place during recognizable time periods over the course of the day. Recognizing a story’s flow through time is essential to understanding the text. In this paper, we try to capture the flow of time through novels by attempting to recognize what time of day each event in the story takes place at. As our motivating example, we use “The Great Gatsby” (Fitzgerald, 1925), a short novel with a fa- miliar plot that can be analyzed with our techniques. Figure 1 presents the work of two human annota- tors, independently making their best guesses of the clock time at every paragraph of the book. The x-axis of Figure 1 represents the full text of the book as enumerated by paragraph numbers, while the y-axis represents the 24 hours of the day. The times identified by the annotators are shown in blue, while our model’s time predictions are shown in red. We first note that there is general but not perfect agreement between the annotators, with an aver- age disagreement of 1.85 hours. There is also a strong general agreement between the model and the annotators, with an average absolute error of 2.62 hours, computed by taking the minimum time difference from either of …

Content truncated to 500 words